Emulsion injections play a vital role in clinical applications, including fat emulsions, total parenteral nutrition (TPN), and anesthetics like propofol. Despite their strong therapeutic effects, their inherently opaque milky appearance makes them one of the most challenging dosage forms to inspect.

Inspection Challenges: Detecting Defects Through a Milky Haze

Compared to clear aqueous injections, emulsions present multiple unique challenges:

• Foreign particles are hard to detect: Due to the opacity of emulsions, traditional transmitted light cannot penetrate the liquid, allowing foreign matter such as glass or metal fragments to remain hidden.
• Bubbles resemble defects: Naturally occurring microbubbles in emulsions often resemble actual defects, making differentiation difficult.
• External interference: During rotary inspection, spots or hairs on the outer wall can be mistaken for internal defects.
• Traditional inspection fails: Manual inspection and general-purpose equipment struggle to detect or accurately judge defects against an opaque white background.

Technical Solution: Integrating Optics and AI to Enable Reliable Emulsion Inspection

To overcome opacity and interference challenges, Truking has developed a complete solution combining optical acquisition with intelligent recognition to detect and classify multiple types of defects in emulsion products.

Our system includes:

• High-resolution color cameras and high-speed rotary mechanisms to enhance foreign particle detection and image capture of motion paths.
• Multi-angle, multi-mode lighting such as side light, angled top light, and infrared transmission to overcome visibility barriers caused by opacity.
• AI-based particle trajectory analysis to identify object movement and behavior, improving discrimination between defects and false positives.
• Multi-angle image capture paths, covering the liquid surface, bottle wall, bottom, and cap to ensure full defect coverage.

Together, these optical and AI technologies enable reliable detection of various defects including black spots, glass shards, oil droplets, and hair.

Defect Detection in Action: Visible and Accurate

Seeing is not enough—accuracy is key. Traditional image-processing algorithms often fail to distinguish between true defects and interference from bubbles, reflections, or wall contamination, leading to misjudgments.

To address this, Truking has developed proprietary AI models that analyze movement trajectories, speed, and morphological changes to exclude most false positives.

On actual production lines, this technology has improved detection accuracy for some typical emulsion products from under 30% to over 80%.

Common Defects and Interferences in Emulsion Inspection:

• Foreign objects on bottle wall: High-res cameras with AI reduce false alarms caused by external contaminants.
  
• 
  
• • Foreign particles at liquid surface: AI tracking helps distinguish black spots from bubbles.
  
  
• 
  
• • Moving particles: Such as glass shards or black specks in motion, identified through dynamic tracking.
  
  
• 
  
• • Bottom-of-bottle inspection: Algorithms separate true defects from bubbles or residue at the base.
  
  
• 
  

• Conclusion

  Emulsions are widely used in critical care and anesthesia, making quality assurance vital. Their opacity and low-contrast imaging of small particles pose a widely recognized technical challenge in the industry.

  Truking continues to tackle this challenge by integrating multi-source lighting, high-frame-rate imaging, and AI recognition—achieving significant results for many common defects. Acknowledging the complexity of emulsion inspection, we remain committed to further optimizing both optics and algorithms.

  With a rigorous technical mindset, we continuously improve our inspection systems to ensure clearer images and more reliable decisions—fortifying the safety and quality of pharmaceutical products.